2031

War is nothing new. Hell, somewhere in the world has been at war for every part of the last century, more or less. But few wars are as damaging to the global economy as that waged between China on one side and China, the United States, and Japan on the other. The unilateral blockade of China by the United States had something to do with this, certainly, but it still doesn't give the full picture. How the hell did a war over one tiny island in the Pacific kill twenty percent of the global GDP?

One word: semiconductors. Those small electronics are the backbone of the modern world--in everything from smartphones to cars to TVs to computers to advanced medical diagnostic equipment to radios. In today's world, it's more advanced than a light bulb, it probably involves semiconductors.

And Taiwan had a stranglehold on the industry. Not on designing the things--that was always more of an American, Japanese, or Korean endeavor--but on making them. The largest semiconductor foundry in Taiwan, TSMC, had revenues more than double that of its greatest competitors in 2021. When combined with other Taiwan-based foundries UMC, PSMC, and VIS, Taiwan's share of global foundry revenue stands at a staggering 63 percent.

More than commanding a dominant share of foundry revenues, Taiwanese firms maintain a dominant position in the production of cutting-edge semiconductors: two thirds of Taiwan's semiconductor production in 2018 was of transistors 65 nanometers or smaller. With TSMC's dominant position in the research and development of advanced semiconductor production techniques (TSMC was the first firm to roll out 3nm production in 2022 and 2nm production in 2024), bolstered by expansive investment from US firms and Apple's decision to use TSMC to manufacture its future 5nm chips, Taiwan's importance in the global tech industry continued to grow.

And then, the crash. The Sino-Taiwanese War had disastrous consequences for Taiwan's semiconductor industry. As the economy transitioned to war footing, the country's production of semiconductors came to a screeching halt. While Taiwanese-owned semiconductor fabs outside of Taiwan and Mainland China (like TSMC's fabs in Singapore, Arizona, and Washington, UMC's fabs in Japan and Singapore, and VIS's fabs in Singapore) kept production high enough to prevent these firms from going bankrupt (with the help of some government aid money, of course), their production, still predominantly located in Taiwan, was still devastated. Add into that China's deliberate targeting of civilian power stations, Taiwan's full mobilization (which included calling most men below 40 to active-duty service), and the effective blockade of and no-fly zone over the island (which was technically broken upon the arrival of the US Navy, but realistically, not many people were keen to try to trade with Taiwan during the conflict), and the semiconductor industry was in a bad place.

When You Come At the King, You'd Best Not Miss

But a bad place is not dead. In the face of the global semiconductor shortage this conflict caused, countries the world over have decided that this is their opportunity to wrest control of the semiconductor industry from Taiwan. The European Union, Iran, Finland, the EAEU, Russia, Vietnam, India, and more have invested billions into risky new semiconductor companies (rather than, for whatever odd reason, preventing key economic sectors in their own countries from collapsing during this economic crisis), hoping to usurp Taiwan's throne.

They will be sorely disappointed. Building up domestic semiconductor production is hard. It is the cutting edge of industry--a sector that requires insanely precise equipment and manufacturing techniques, highly-trained and educated workers, and massive capital investments. These are not the sort of thing countries can build up overnight. These are the sorts of things that take decades to establish.

Let's take China as an example. For decades, one of China's primary industrial policies has been the development of a domestic semiconductor industry. China is heavily reliant on foreign-built semiconductors, which are the country's second largest foreign import (behind crude oil). Hoping to counteract this, in 2015 Beijing announced its new Made in China 2025 policy initiative which, among other things, sought to increase the amount of semiconductors made in China from about 15 percent in 2015 to 30 percent by 2020 and 70 percent by 2025.

This policy was, by any objective metric, an utter failure. China did manage to increase the number of semiconductors produced in China: according to semiconductor market research firm ICInsights, China's domestic semiconductor production almost doubled between 2014 and 2019--from ~12b USD to ~20b USD. The problem is that the market for semiconductors in China also nearly doubled--from 77b USD in 2014 to 125b USD in 2019. On net, this means that in five years, China only managed a measly 0.5 percentage point increase in the share of its semiconductor demand met by production in China, increasing from 15.1 percent in 2014 to 15.6 percent in 2019. Worse still, half of the semiconductors produced in China were made by foreign companies--often Taiwanese firms, like TSMC and UMC! Projecting these trends out for the rest of Made in China 2025 period is similarly bleak, with domestic production (though doubling again to ~40b USD) only accounting for 20.6 percent of market share.

Keep in mind: prior to the war with Taiwan, China was one of the wealthiest nations on Earth, with one of the most advanced tech sectors outside of the Global North. The global economy was growing at near-unprecedented rates. And after a decade of work, they increased their domestic market share by five percent. What do you think will happen to all of these new semiconductor ventures around the world, attempting to develop brand new manufacturing techniques and establish industrial-scale production of cutting-edge semiconductor from scratch--all in the middle of the worst economic crisis in over a century? The answer is, it'll probably go pretty damn poorly.

The Reports of My Death Are Greatly Exaggerated

Maybe these lofty goals would be more attainable if the Taiwanese semiconductor industry had been totally and utterly destroyed. Many parts of Taiwan were leveled. Kinmen County, under orders from the PLA to "return the island to the Stone Age", saw a sixth of its civilian population killed and another fifth wounded. Lienchiang County, when it finally fell to the Chinese in September 2029, saw its civilian population of roughly 13,000 summarily executed by the landing force. Vast swathes of Taipei were damaged or destroyed (though the worst of it was averted by the brave pilots of the ROCAF) and parts of Taichung City came under artillery fire following direct orders from the PLA. If any of these had been the hubs of Taiwan's semiconductor industry, then perhaps these new ventures across the globe would be successful.

Fortunately for Taiwan (and unfortunately for our newfound competitors), they were not. Taiwan's semiconductor industry is heavily concentrated in Hsinchu Sciences Park, straddling the border between Hsinchu City and Hsinchu County. UMC's Taiwanese production, for example, is all located in Hsinchu (with the exception of one fab in Tainan), as is TSMC's (though they're a bit more diversified, with fabs in Tainan, Taoyuan City, and Taichung). Hsinchu, well away from the targeted civilian centers of Taichung and Taipei, was never subjected to any sort of concentrated terror bombing campaign. Sure, military targets and power stations were hit. A few highways probably got bombed. But for the most part, the actual fabs themselves, and the outrageously expensive equipment stationed within them, are still intact. All they need is some capital, some workers, some inputs (electricity, minerals, blank wafers), and a market to sell to (and by God, in this global crisis, is there a market to sell to!), and they're back to cranking out semiconductors like the war never even happened. Well, maybe with some hit to production, but you get the point: Taiwan's semiconductor industry will be back up to almost pre-war production levels long before the new fabs abroad are finished.

Making matters worse, making new semiconductor ventures and factories is insanely expensive, costing billions of dollars--TSMC's first 3nm fab, for example, cost a staggering 19.5b USD. This, coupled with greater-than-average volatility, has contributed to a massive consolidation at the leading-edge of the industry: while almost 30 firms were able to produce cutting-edge semiconductors in 2001 (130nm at the time, by 2020, only three firms (Intel, Samsung, TSMC) could produce cutting-edge chips (5nm, 26 times smaller than the cutting-edge chips of 2001)--and signs point to that number dropping even lower, as Intel considers outsourcing its advanced chip manufacturing to TSMC.

So, let's compare the option facing global investors right now. On the one hand, they can invest in one of the many new semiconductor firms popping up throughout the globe. The startup costs will be immense (research and development of the necessary production techniques, construction of the fabs to make the semiconductors, procurement of the expensive, high-tech equipment necessary to make the finished product). There will be no guarantee of a return on their investment. The firm may fail to develop the necessary production techniques. It may fail to develop its own chips (if it's an integrated manufacturer or fabless manufacturer) or fail to secure contracts to produce the chips of other firms (if it's a foundry). And even if it manages to get all of those right: semiconductor prices are high right now, but will the firm be able to survive when they drop back to pre-war levels? Below pre-war levels, even, as the increased supply lowers prices? Failing at any step of this process will result in massive losses in the investment.

Or, they can invest in Taiwan-based semiconductor manufacturing. They don't need to build new fabs; they already exist, and were never destroyed. They don't need to design new production methods; they are already at the cutting edge of the market (though maybe Samsung has caught up on TSMC's slight lead at this point). They don't need to secure new contracts: they're already to go-to producer for hundreds of electronics companies. And, perhaps most importantly, they've proven that they can operate at the previous market-clearing levels. What disadvantages once existed in Taiwan's production (namely, the looming threat of Chinese invasion) have largely been dispelled following the signing of the Treaty of Singapore and Beijing's full recognition of Taiwan as a sovereign and independent nation.

So why, if you're a private investor, would you pick anywhere other than Taiwan?

Making Their Dollar Work for You

Looking abroad, Taiwanese semiconductor firms are squealing with glee. If everyone wants to spend money on building up a domestic semiconductor industry, why not take advantage? Several firms, like TSMC and UMC, have sizeable fabs in China that they would love to relocate elsewhere--both because doing business in China is very unpopular in Taiwan these days, and because Taiwan's withdrawal from the Economic Cooperation Framework Agreement has shifted the financial calculus of doing business in China. Besides: more money and more market share abroad (both of which can be gained by moving abroad) is never a bad thing.

TSMC and UMC have started to put feelers out to nations subsidizing domestic semiconductor production--namely the European Union and India--to discuss the possibility of relocating some of its production currently located in China in exchange for access to the government tax credits and subsidies being offered to the semiconductor industry. This is viewed as a win-win exchange by these firms: they get access to government subsidies and tax incentives and access to new markets, while the sponsoring nations get access to domestic (albeit foreign-owned) semiconductor production by a proven market leader at a substantially lower cost and risk than going at it alone.

Now, the firms aren't threatening this in any way, but it should be kept in mind that under the WTO Agreement on Subsidies and Countervailing Measures, these governments--which are all WTO members--kind of have to make these benefits available to them, too. Otherwise, it's an actionable subsidy. And the major semiconductor producers of the world--Taiwan, Japan, South Korea, and the United States--are known to be a little trigger-happy on such matters.

Safeguarding Taiwanese Production

Given the strategic importance of Taiwan's semiconductor industry (semiconductor exports are the largest individual component of the Taiwanese economy--to say nothing of the knock-on effects the industry has on the economy), restoring Taiwan's semiconductor production capabilities is the nation's top priority. Under the newly-passed Protecting the Industry of Taiwan (PIT) Act, Taiwan has opened a massive line of low-interest credit--upwards of 25b USD--to firms operating in Taiwan's semiconductor industry (this credit line is not limited to firms headquartered in Taiwan, but recipient firms do have to have some operations there, and the credit available to them scales based on the size of their operations), intended to protect the industry from outright bankruptcy during this economic crash. Taiwan has also demobilized all reservists who were previously employed in Taiwan's semiconductor industry (most reservists have been demobilized since the signing of the Treaty of Singapore, save those still assisting in relief and rebuilding efforts, but these ones in particular have been fully demobilized).

In fact, Taiwan's universal conscription, as well as the Chinese blockade and no-fly zone over the island, has provided an unintended benefit to Taiwan's efforts to maintain its semiconductor industry. Taiwan's reserves, numbering over two million strong, includes most men between the ages of 20 and 40--one of the key demographics employed in Taiwan's tech industry. Since they were in uniform, it was considerably more difficult for them to flee the conflict as a refugee (since this would be desertion) than it was for Taiwan's civilian population--and it was already pretty hard for them, considering most of the island's ports were blockaded and civilian aviation was effectively shut down for the entirety of the conflict. Which means that most of the people employed in Taiwan's semiconductor industry--all of that precious human capital--are still in the country.

Under the PIT Act, Taiwan's semiconductor industry has also been selected for special preference in the reconstruction effort. The industry is one of the main consumers of electricity on Taiwan: TSMC alone was responsible for 7.2 percent of Taiwan's total power consumption in 2022, making it far and away the largest single energy consumer in the country. Unfortunately, electricity is in extremely short supply in Taiwan at the moment due to China's bombing campaign against civilian power stations. Strict energy rationing is in effect throughout the island, with certain sections of every city on receiving power during certain parts of the day until grid capacity can be fully restored. The only sectors exempted from this are the medical sector, public transportation (subways and Taiwan's HSR, for instance), the military (which is mostly operating on its own generators at this point anyway), critical government functions (wouldn't want the Legislative Yuan to be without power!), and firms deemed by the government to be of critical importance, which includes all firms, foreign and domestic, involved in Taiwan's semiconductor industry. Furthermore, the country has locked in energy prices charged by state-owned electricity firmed Taipower. This runs up Taipower's deficit some more, but that seems like a very small problem in the face of total economic collapse.